Video is by far the predominant type of traffic by volume in the Mobile networks (the video traffic load in a typical RNC can be about 40% of the total volume, from which 61% of HTTP traffic is video contribution to volume, which in turn equates to 40% of all traffic since HTTP forms 66% of all traffic). The largest number of the traffic video downloads is carried out via HTTP progressive downloading (the video file is actually downloaded to the end users equipment, for example, through a browser such as YouTube). The remaining video downloads use streaming media servers.
On the other hand, it has been estimated that one in four users cancels a video stream early after initiating its download. Thus, considerable RAN resources are wasted in delivering the video, since an amount of video is already downloaded and buffered in advance of the point in the stream at which the user aborts the viewing session.
Moreover, the growth rate of data traffic (including video) is increasing twofold in less than one year. However, the revenues coming from the data traffic are not rising at the same tempo, which leads to the necessity to control the data traffic based on network resources.
The application specific traffic optimization in a wireless link disclosed by US 2002/0118649 is mainly concerned of detecting the data type of the packets carried on a wireless connection by examining their link performance characteristics, such as a port number, and with the assessment of data types, in the case of audio and video, certain link control parameters that govern the transport of the packets are determined and applied to. The computation of said link control parameters is based on fixed preconfigured values, a mere determination of a particular, static, packet flow model as a consequence of the classification of the data.